1

The Effect of Listing Price Strategy on Real Estate

Negotiations: An Experimental Study

Eric Cardella1 Michael Seiler

2

Texas Tech University The College of William and Mary

February 1, 2014

Comments Welcome

Abstract

When selling a home, an important decision for the home owner is choosing an optimal listing

price. This decision will depend in large part on how the chosen list price impacts the post

negotiation final sale price of the home. In this study, we design an experiment that enables us to

identify how different types of common list price strategies affect housing negotiations.

Specifically, we examine how rounded, just below and precise list prices impact the negotiation

behavior of the buyer and seller and, ultimately, the final sale price of the home. Our results

indicate that the initial list price does play an important role in the negotiation process.

Interestingly, these impacts generally attenuate with negotiating experience.

Acknowledgements: This study benefited greatly from contributions made by Xudong An,

Marlon Boarnet, Dennis Capozza, Thomas Davidoff, Jeff Fisher, Donald Haurin, Patrick

Hendershott, Christian Hilber, David Ling, Steve Malpezzi, Norm Miller, Bill Mundy, and Jeff

Zabel. All errors remain our own.

1 Rawls College of Business, Texas Tech University, Lubbock, TX 79409; Telephone: (858) 395-6699; Email:

eric.cardella@ttu.edu. 2 Mason School of Business, The College of William & Mary, P.O. Box 8795, Williamsburg, VA 23187-8795;

Telephone: (757) 585-6100; Email: Michael.Seiler@mason.wm.edu

2

1 Introduction

A key decision in selling a home is selecting the listing price. Typically, this initial list price

serves as the starting point in the negotiation process, which ultimately impacts the final sale

price. Given the magnitude of house prices, even a small percentage change in the final sale price

can have significant financial ramifications for the seller. As a result, one of the chief aims of the

seller is choosing the optimal list price. The purpose of this study is to answer the question,

“How do initial list prices impact negotiations and, ultimately, the final sale price?”

In this study, we develop a novel experimental design that enables us to systematically explore

the effect of the initial list price strategy on purchase negotiations. Specially, we study how

variation in the thousands digit of the list price effects the negotiation behavior of the buyer and

seller and, consequently, the final sale price. We consider three different types of listing prices,

which we refer to as list price strategies: (i) round, (ii) just below, and (iii) precise. Consistent

with the characterization in Beracha & Seiler (2014), a “round” price is one where the thousands

digit is either 0 or 5, “just below” is a price where the thousands digit is either 4 or 9, and a

“precise” price is defined as having a thousands digit of either 1, 2, 3, 6, 7, or 8.

Within the experiment, participants engage in a stylized, bi-lateral housing negotiation. As part

of the experimental design, we exogenously vary the list price strategy across treatments, while

holding all other aspects of the negotiation process constant. In the “rounded” (R) treatment the

list price is set to $200,000, in the “just below” (JB) treatment the list price is set to $199,000.

For “precise” pricing, we consider two variations: the high precise (HP) treatment features a list

price of $201,326, whereas the low precise (LP) treatment features a list price of $198,674. This

systematic manipulation of the list price strategy enables us to compare outcomes across

treatments to identify the corresponding effect of the list price strategy. Furthermore, we are able

to observe all intermediate steps within the negotiation process including the entire sequence of

offers and counter-offers, which allows us to identify how the various list price strategies

separately affect the negotiation behavior of buyers and sellers. This level of investigation is not

possible using transactions data because with transactions only the outcome is observable, not

the intermediate steps taken to achieve the outcome.

A small body of literature exists that focuses on the relation between types of list prices and real

estate sales. Allen & Dare (2004) use transactions data from Florida and document that “charm”

pricing is associated with higher final sale prices.3 Palmon et al. (2004) use transactions data

from Texas and find that “just below even” list prices are associated with lower final sale prices,

compared to “even” list prices.4 The most closely related study to ours is a recent paper by

3 Allen & Dare (2004) define a “charm” price as a price ending in 500, 900, 4,900, 5,000, 9,000, or any other

non-zero number. Hence, their characterization of a charm price does not match either of the three list price strategies we consider, and it encompasses (as a subset) what we define as just below and precise. Furthermore, their broad characterization of a charm price includes variation in all digits up to the thousands digit. Therefore, there is no clear way to connect the results from Allen and Dare to our results.

4 Palmon et al. (2004) characterize “even” prices as those ending with 000, and “just below even” as those

ending with 900. Thus, their definitions would be along the lines of our definitions of round and just below pricing.

3

Beracha & Seiler (2014) who use data on over 300,000 residential real estate transactions in

Hampton Roads, Virginia, spanning 1993 – 2011. The authors find that just below pricing is

associated with the largest negotiated discounts. However, because just below pricing is also

associated with the greatest degree of over-pricing by sellers, just below pricing leads to the

greatest net proceeds to the seller. Our study is intended to complement this prior work by

providing a much deeper understanding of the previously unobserved negotiation process that

leads to the final sale price, which we achieve through an experimental design. There is also a

body of literature looking more generally at how list price magnitudes impact various aspect of

real estate transactions; we review this literature, as well as some of the literature on consumer

good pricing, in the next section.

We contend that a controlled laboratory setting affords us several advantages relative to these

prior transactions-based empirical studies. Firstly, in the lab we are able to induce, and therefore

observe, both the buyer’s and seller’s reservation values, which are unobservable when using

transactions data. Secondly, we are able to systematically control for the fundamental value of

the property being transacted, which is latent for real properties. This mitigates the endogeneity

problem resulting from systematic over/under-pricing associated with certain list price strategies

(e.g., Beracha & Seiler, 2014 who document over-pricing of homes listed at a just below price).

Thirdly, we are able to abstract from the dimension of housing quality. This is important because

with transactions data, housing quality is either imprecisely measured or not measured at all.

This leads to selection bias when estimating the effect of list prices on final sale prices.

Alternatively stated, sellers with lower or higher quality properties may systematically gravitate

towards certain list price strategies. Paloma et al. (2004) point to such a concern by noting

“Selection bias might be present if sellers’ decision to list their properties at even or just-below-

even prices is associated with some unobserved variables that are correlated with the inherent

value of the property” (p. 129). Because the underlying quality of the property is unobservable in

the field, it is not possible using transactions data to separate the effect of list prices on final

transaction prices from possible differences in property quality, which are endogenously

associated with list price. This problem is overcome in an experimental setting.

Finally, by randomly and exogenously assigning list price strategies within a negotiating dyad,

we are able to avoid possible selection bias based on the negotiating style/ability of sellers. There

is likely to be substantial heterogeneity among sellers in the housing market, and sellers of either

low or high negotiation ability may systematically gravitate toward certain list price strategies.

The endogenous selection of list price strategy by sellers in transactions data, again, makes it

difficult to isolate the effect of list prices on final sale prices. However, by randomly and

exogenously assigning sellers to list price strategies, we can cleanly identify (via average

treatment effects) the impact of list pricing strategies.

Overall, our results indicate that the list price strategy does impact negotiations. In particular, we

find that high precise pricing leads to the highest final sale price. With regard to negotiated

respectively. However, as these authors examine variations in the hundreds digit, their results are not directly comparable to ours.

4

discount, high precise pricing results in the smallest percentage discount while just below pricing

results in the largest percentage discount, relative to the list price. Our data also suggest that list

prices affect buyers and sellers differently. Namely, buyers make the lowest offers when the

price is rounded, while sellers make the lowest counter-offers when the price is just below.

Furthermore, the effect of the list price has a more persistent effect on sellers throughout the

negotiation process. Lastly, we find that negotiating experience tends to attenuate the observed

effects of the various list price strategies.

2 Related Literature

While the primary motivation of this study is to investigate list price strategies in residential real

estate transactions, there does exist a mature body of literature relating, more generally, to list

prices in consumer goods markets. In particular, the presence of a disproportionate number of

prices ending in 9 has been empirically documented (Twedt, 1965; Friedman, 1967; Schindler &

Wiman, 1989; Kreul, 1992; Shindler & Kirby, 1997; Holdershaw et al., 1997).5 Furthermore,

several studies have documented that prices ending in 9 can be effective in generating higher

sales (Schindler & Kibarian, 1996; Stiving & Winer, 1997; Gendall et al., 1997; Kalyanam &

Shively, 1998; Gednek & Sattler, 1999; Anderson & Simester, 2003; and Bizer & Schindler,

2005). The idea is that consumers tend to drop-off the right most digits (Basu, 1997, 2007; Bizer

& Schindler, 2005) or that there is a left-digit bias (Poltrock & Schwartz, 1984; Korvost &

Damian, 2008); hence, consumers pay less attention to those rightmost digits.6

While these prior studies suggest list price strategies may influence residential real estate

transactions, they may not translate directly. The reason is that in real estate transactions, unlike

many consumer goods markets, the list price is generally viewed as negotiable. That is, the list

price serves as the starting point of the negotiations. Furthermore, purchase prices in real estate

transactions tend to be much larger than most other retail consumer goods. As a result, list price

strategies may impact the negotiation process between the buyer and seller differently than just

the buyer’s decision about whether to purchase a good at a fixed posted price. That said, we view

this study as contributing to the extant literature on list price strategies by investigating their

effect on negotiated outcomes in real estate transactions where prices are negotiable.

With regard to real estate transactions, there is a body of literature on the relation between list

prices and sale prices (Anglin et al., 2003; Han & Strange, 2014; Haurin et al., 2010; Haurin et

5 The phenomenon of prices ending in 9 have been referred to and explored under various terms including

pricing in the nines (Basu, 1997; 2007), 9-ending pricing (Gednek & Sattler, 1999; Bizer & Schindler, 2005), odd-pricing (Monroe, 1979; Evans & Berman, 1997), and just below pricing (Stiving & Winer, 1997; Beracha & Seiler, 2014).

6 Formal models of inattention bias have been developed by Chetty et al. (2009), Della Vigna (2009), and

Finkelstein (2009). Lacetera et al. (2012) provide empirical evidence of inattention bias with respect to rightmost digits using used car transaction data; specifically, the authors find that there are discontinuous sales price drops at 10,000 and 1,000 mile odometer thresholds.

5

al., 2013; Knight, 2002; and Miller & Sklarz, 1987)7. However, these studies focus primarily on

how the relative position of the list price to the fundamental property value impact final sale

prices and time on the market (as opposed to examining the type of list price strategy employed).

There are a few prior empirical studies that look specifically at list price strategies (Allen &

Dare, 2004; Paloma et al., 2004; and Beracha & Seiler, 2014). However, as discussed in the

introduction, these empirical studies using transactions data provide challenges in identifying the

causal impact of list price strategies. By using a controlled experiment where we systematically

and randomly vary list price strategies, we are able to causally identify how the list price impacts

both the negotiation process and the final sale price. Thus, we see this study as an important and

necessary step in deepening our understanding how list prices impact real estate negotiations and

a contribution to the extant empirical literature.

Real estate in not the only market where prices are generally viewed as negotiable. Other

examples include: automobile, big ticket consumer durables, services, and used goods (see

Cardella, 2014 for a discussion and review of the literature relating more generally to flexible

price markets). An important issue in flexible price markets is whether it is profitable for sellers

to implement a pricing scheme where buyers can (attempt) to negotiate, or rather, should sellers

commit to a fixed, non-negotiable price? Several papers have theoretically explored this topic

(Bester, 1993, 1994; Wang, 1995; Arnold & Lippman, 1998; Desai & Purohit, 2004; Raskovich,

2007; Gill & Thanassoulis, 2009, 2013). Often, the optimal pricing decisions for sellers and the

equilibrium predictions depend, in part, on the negotiation behavior of the buyers. Thus, in terms

using these models to generate predictions about optimal seller behavior and market outcomes, it

is important to understand the bargaining tendencies of buyers. However, these models abstract

away from the possibility that the negotiation behavior of buyers may be endogenous to the

chosen flexible list price. The results from this study regarding how list price strategies impact

negotiations can have important implications in other markets regarding whether sellers should

allow prices to be flexible and if so, what is the optimal list price.

3 Experimental Design

Our experimental design is centered around a stylized, bi-lateral real estate negotiation. In the

negotiation, there is a seller who is endowed with a fictitious house that is for sale at an initial list

price, denoted . The other party is a buyer, who is tasked with trying to negotiate the purchase

of the house from the seller.

The negotiation process begins with the buyer choosing to either purchase the house at the initial

list price, or make a counter-offer to the seller. Conditional on the buyer making a counter-offer,

the seller then has an opportunity to either accept the buyer’s counter-offer, or reject and make a

counter-offer back to the buyer. If the seller accepts, then the negotiation ends. If the seller

makes a counter-offer, the buyer then has an opportunity to accept the counter-offer, or reject

7 Anenberg & Kung (2014) examine list price changes when nearby REO properties enter the market. Anenberg

& Laufer (2014) use the spread of list price to sale price ratio as a way to create forward-looking price home indexes. Neither study approaches the level of granularity in, or focus of, our study.

6

and make a counter-offer. The negotiation continues in this alternate-offer style format until

either: (i) the buyer or seller accepts the offer of the other party, or (ii) the maximum time

allotment of 3-minutes elapses without an agreement (in which case the house is not transacted).

Sellers are privately informed of their reservation value, denoted by r, for the house at the

beginning of the negotiation process. Similarly, buyers are privately informed at the beginning of

the negotiations of their value for the house, denoted by . If the buyer and seller are able to

reach a negotiated agreement, then denotes the final negotiated price. In terms of payoffs,

sellers are privately informed that they will receive a payoff of if they sell the house, and

a payoff of zero if they fail to sell the house. Buyers are privately informed that they will receive

a payoff of if they buy the house, and a payoff of zero if they fail to buy the house. To

remain as close as possible to actual residential real estate prices, and to allow for variation in the

thousands digit of the list price, all monetary values are in terms of 6-figure experimental

currency dollars (ECD), which are then converted into actual dollars at the conclusion of the

experiment.

The main motivation of this study is to investigate how the listing price of the house impacts the

negotiation process and the final transaction price. To do so, we consider 4 different list prices

for the house, which are characterized as follows:

Rounded Price (R):

Just Below Price (JB):

High Precise Price (HP):

Low Precise Price (LP):

Consistent with the characterization used by Beracha & Seiler (2014), the rounded price features

a thousands digit of 0, the just below price a has thousands digit of 9, and the two precise list

prices feature thousands digits of 1 and 8.

Because a precise list price can be either above or below the rounded price, it is possible that

these two different types of precise prices have different effects on the negotiation. By

considering two different precise prices, one that is above the rounded price (HP) and one that is

symmetrically below the rounded price (LP), we can test if and to what extent the position of the

precise price matters. Furthermore, it allows us to rule out the possibility that any observed

difference in the precise price is merely a result of it being the lowest or highest price. This will

provide a more robust analysis of the effect of precise pricing than that of Beracha & Seiler

(2014) who essentially lump both low and high precise pricing into one category.

In addition to variation in the list price, another feature that is likely to characterize real estate

negotiations is variation in the price originally paid by the seller.8 It is possible that this variation

in the original purchase price of the house may mediate the effect of list price strategy. In

8 One source of this variation in the price paid could be the result of the length of time the seller has owned the

house; a seller who has owned a house for a longer period of time will have likely purchased the house at a lower price than a seller of a comparable home who has owned the house for a much shorter period of time.

7

particular, if the list price of the home is close to the purchase price, then the purchase price may

act as an induced anchor or “price floor” during the negotiations, from the perspective of the

seller (see Kahneman, 1992 for a discussion of anchor points in negotiations and Kristensen &

Gärling, 2000 for evidence that anchor points impact negotiations. In this case, sellers may be

resistant to sell the house at a price below the original price they paid for the house, i.e., these

sellers may be resistant to sell at a price that generates a “perceived” loss. Thus, the seller may be

less inclined to negotiate when the original purchase price is closer to the list price.9

Consequently, this rigidity by the seller could then attenuate any possible effects that variation in

the list price strategies may have on negotiations.

To explore whether the original purchase price paid by the seller impacts negotiations, and

possibly interacts with the list price strategy, we consider two different purchase prices for the

home, denoted by . In the high purchase price case, sellers are informed, prior to the start of

the negotiation, that they originally paid a price of for the house. In the other

case, the low purchase price, sellers are informed that they originally paid for

the house. The idea being that in the high price case the purchase price is set to a value that is

slightly lower than the list price; thus, sellers in this case who do not want to sell the house for a

perceived loss would not want to negotiate below $195,000. In the low price case, sellers who do

not want to sell the house for a perceived loss would not want to negotiate below $185,000.

Therefore, all final negotiated prices between $185,000 and $195,000 would be perceived by

sellers as a loss in the high purchase price case and as a gain in the low purchase price case.

Hence, this “break even” constraint puts more upward pressure on the price in the high purchase

price case compared to the low purchase price case.

To summarize, we consider four different list price strategies interacted with two variations in

the seller’s initial purchase price, resulting in a 4 X 2 experimental design. Table 1 displays the

full list of parameter values used in each of the 8 different conditions. From Table 1, we see that

in all 8 conditions, the buyer’s value, , is held constant at $205,000.10

In addition, the seller’s

reservation, , is always set to be $20,000 below the list price. This ensures that the bargaining

zone, which is the difference between the list price and the seller’s reservation value, is constant

at $20,000 for each of the four list price conditions. Thus, by comparing negotiation behavior

and the resulting negotiation outcome across the four different list price conditions, we are able

to identify how the list price strategy impacts negotiations.

[Insert Table 1 – Parameter Value Combinations by Treatment]

9 The idea that agents may behave differently based on whether outcomes are in the domain or perceived

gains/losses was first introduced in Prospect Theory by Kahneman & Tversky (1979). In the model, an agent’s value function is S-shaped around the reference point; convex in the domain of losses and concave in the domain of gains, which induces more (less) risk seeking in the domain of losses (gains). In the context of negotiations, this differential behavior can then impact counter-offers, concession behavior, and the likelihood that an agreement is reached (see Neale & Bazerman, 1991 and Kahneman, 1992 for a though discussion and a review of the literature documenting the impact of variation in the gain/loss domain on negotiations.

10 Because the buyer’s value is higher than the list price, it is pareto efficient for the house to be transacted at

any price between the list price and the seller’s reservation value.

8

The only difference between the four conditions in Panels A and B is whether the seller was told

that the original purchase price was high at or low at . Note that

the seller’s payoff does not depend on the original purchase price, only the seller’s reservation

value. So, manipulating the information the seller gets regarding the original purchase price does

not impact the financial incentives in any way. However, this manipulation does alter the

reference point for the sellers around which they evaluate perceived gains and losses, which is its

intended purpose. Therefore, by comparing negotiation behavior and outcomes across the

conditions in Panels A and B (holding constant the list price, but varying the original purchase

price), we can identify if the original purchase price impacts negotiations, and to what extent this

may interact with different list price strategies.

We acknowledge that the negotiation process we implement in our design is stylized and may

abstract away from features that are present in real house negotiations. That said, the stylized

features and the controlled laboratory setting enable us to shed light on the causal effect of

variation in the list price on negotiations. In particular, by systematically varying and

exogenously assigning the list price in the negotiation, as opposed to the seller selecting the list

price, we circumvent one possible source of selection bias based on the seller’s attitude toward or

ability in negotiating. In addition, we abstract from any dimension of house quality, which also

eliminates possible bias in the effect of list price strategies based on unobserved house quality.

Lastly, because we implement a controlled laboratory experiment we are able to observe all the

intermediate steps of the negotiation process (i.e., the sequence of counter-offers from each

party). This allows us to more rigorously examine how the list price strategy separately impacts

the behavior of the buyer and the seller, as well as the final outcome of the negotiations.

4 Experimental Procedure

All experimental sessions were conducted in the Economics Science Laboratory (ESL) at the

University of Arizona in October 2013. The experiment was computerized and the software was

programmed using z-Tree.11

All interactions between subjects were anonymous. The subject

pool consisted of undergraduate students who were recruited via email from an established

database. We ran 8 sessions, and a total of 132 subjects (66 negotiating pairs) participated in the

experiment. In each session, subjects were randomly assigned to either the role of buyer or the

role of seller and remained in this role for the entire experiment. A copy of the participant

instructions can be found in the appendix.

We implemented a combination of a within and between subjects design. The manipulation of

the high/low original purchase price conditions was done between groups: four sessions of the

high purchase price and four sessions of the low purchase price. The manipulation of the four list

price strategies was done within groups; in each experimental session, subjects participated in

four separate negotiations corresponding to each of the four different list price strategy

treatments. To help mitigate repeated game effects, all subjects were randomly re-matched with a

11

We thank Urs Fischbacher (Fischbacher, 2007) for providing this software.

9

different subject of the opposite role between each negotiation, and subjects were clearly

informed of this matching protocol via the instructions.

To control for possible order effects of the within subjects manipulation of list price treatments,

we used a balanced design with four different orders of the treatments. Each of the four different

list price treatments appeared as the first round of negotiations in one of the orderings.12

That is,

approximately ¼ of the subjects participated in each of the four list price treatments the first

round of negotiations. This balanced design enables us to test for treatment effects along the

dimension of list prices in the first round of negotiation only, without any possible confounds. At

the same time, we are able to test for treatment effects in the aggregate across all four rounds of

negotiations, and explore whether the (possible) treatment effects are attenuated as subjects gain

experience.

After completing the four rounds of negotiations, subjects were asked to complete a short

questionnaire (about 5 minutes in length). The purpose of the questionnaire was to elicit some

subject level information on demographics, experience and attitude toward negotiations, and

personality traits. We then use this information to explore possible heterogeneous treatments

effects. More detail on the information we elicit and the specific questions we use is provided in

in the Results section, where we analyze the effect of these factors.

In terms of compensation, subjects were randomly paid for one of the four rounds of negotiation,

and they were clearly informed of this payment scheme within the instructions. The random

payment for one round was intended to mitigate possible confounding wealth and hedging

effects. Upon completion of all four rounds of negotiations and the questionnaire, subjects

individually rolled a 4-sided die to determine which round was selected for payment. Their

earnings from the selected round where then converted at a rate of 1000 ECDs = $1. In addition,

all subjects received a $5 show-up payment. On average, subjects earned a total of $17.50, and

each session lasted approximately 45 minutes.

5 Results

We proceed by presenting the aggregate experimental results. We first look at the data

concerning the effect of the original purchase price paid by the seller. We then look at the data

concerning the effect of the four different list price strategies. We conclude exploring some

possible heterogeneous treatment effects based on subject level demographics, negotiation

experience, and personality traits. The main findings are summarized along the way, while we

postpone a formal discussion of the results until Section 6.

12

With 4 different treatments, it was not practical to run 24 sessions corresponding to the 24 unique orderings of the 4 treatments. As an alternative, we ran 4 different orderings with each treatment coming first in one of the orderings followed by a random assignment of the other 3 treatments. This resulted in the following 4 different orderings of the treatments: (i) R-JB-LP-HP, (ii) JB-LP-HP-R, (iii) LP-JB-R-HP, (iv) HP-LP-JB-R.

10

5.1 The Effect of Original Purchase Price Paid by the Seller

To test whether the original purchase price paid by the seller (i.e., whether the sale of the house

is in the perceived loss or gain domain) has an effect on seller behavior and the negotiation

outcome, we compare the data across the high ($195,000) and low ($185,000) purchase price

conditions, aggregated over the four different orderings of the list price strategies. In particular,

for the negotiation outcome we look at the final sale price (which is the final negotiated price,

conditional on agreement between buyer and seller). For seller behavior, we also look at the

seller’s 1st counter-offer, which is the price the seller counter-offered back to the buyer

(conditional on the buyer making an initial offer to the seller). Table 2 compares the final sale

price and the seller’s 1st counter-offer across the two purchase price conditions for each of the

four different list price strategies.

[Insert Table 2 – The Effect of the Purchase Price Paid by Seller]

Looking first at the final sale price (Panel A of Table 2), we see that when the list price is HP, R

and JB, there are no significant differences in the final sale price between the high and low

purchase price conditions (p = .613, p = .443, and p = .714, respectively). With a LP list price,

the data does reveal that the final sale price is significantly higher in the high purchase price

condition (p = .045). However, when taken as a group, an ANOVA test does not reveal a

significant effect of the initial purchase price on the final sale price (p = .640).

When considering the seller’s 1st counter-offer (Panel B of Table 2), a similar pattern emerges in

the data. Specifically, there is no difference in the seller’s 1st counter-offer when the list price is

HP, R, or JB (p = .374, p = .113, and p = .563, respectively). But, when the list price is LP, the

seller’s 1st counter-offer is marginally significantly higher in the high purchase price condition (p

= .075). Again, when examined across all list price strategies, an ANOVA test does not reveal a

significant effect based on the initial purchase price (p = .174).

Taken together, the data reveals that the manipulation of original purchase price seems to have

had little impact on seller behavior and the negotiation outcome. In particular, when examining

pair-wise comparisons across the two purchase price conditions, the difference is only significant

(marginally in the case of the seller’s 1st counter-offer) when the list price is LP. Furthermore,

the initial purchase price is not a significant factor when comparing across all four list price

strategies. These results imply that sellers are behaving rationally by ignoring false reference

points (i.e., the initial purchase price).

These findings are summarized below:

Result 1: The negotiation behavior of sellers and the final negotiation outcome is not impacted

by whether the original purchase price paid by the seller is high ($195,000) or low ($185,000).

Because there was little difference based on the original purchase price, there is little reason to

expect the original purchase price condition to differentially interact with the four list price

treatments. Thus, for the remainder of the analysis we pool the data from the high and low

purchase price conditions.

11

5.2 The Effect of List Price Strategy on Negotiations

Next we turn to the primary focus of the study, the effect of list price strategy on negotiations.

When comparing across list price strategies, we present and analyze the data from just the first

round (which consists of only the subset of subjects who faced that given list price strategy in the

first negotiating round), as well as the aggregate data (which consist of all subjects in the

sample). We first look at how the list price strategy impacts the final sale price, followed by how

the list price strategy separately impacts the negotiation behavior of buyers and sellers.

5.2.1 The Effect of List Price Strategy on Final Sale Price and Negotiation Outcome

Figure 1 displays the average final sale price for each of the four list price strategies. For only the

first round, the average final sale prices when the list price was HP, R, JB, and LP are $193,288,

$189,763, $187,617, and $187,967, respectively. For the aggregate data, the average final sale

prices when the list price was HP, R, JB, and LP are $191,598, $190.203, $188,754, and

188,490, respectively. The pattern that emerges is a strong and significant correlation between

the list price and the final sale price; the lower the initial list price, the lower the final price.

Specifically, a Jonckheere-Terpstra non-parametric test for ordered alternatives strongly rejects

the null of equality in favor of the descending ordered alternative for both the first round (p =

.003) and aggregate data (p < .001).

[Insert Figure 1: Comparison of Final Sale Price by List Price Strategy]

While the correlation between list price and final sale price is interesting, it is not fully indicative

of how the list price strategy affects the final sale price. To better measure how the list price

strategy impacts the final price, we calculate the % discount, which is the percentage difference

between the list price and the final sale price. Table 3 compares the average % discount for each

of the four list price strategies, in the first round and in the aggregate data. Considering only the

first round, we see that the average % discount when the list prices are HP, R. JB, and LP are

3.94%, 5.12%, 5.72%, and 5.39%, respectively. While a Kruskal-Wallis test fails to reject the

null of equality of these four samples (p = .309), a pairwise comparison between the % discount

when the list price is HP and JB using a Mann-Whitney test reveals a significant difference (p =

.056). Thus, in the first round, HP pricing yields the smallest % discount while JB pricing leads

to the largest % discount.

[Insert Table 3 – Effect of List Price Strategy on Negotiated Discount]

When looking at the aggregate data, the corresponding average % discounts are 4.83%, 4.90%,

5.15%, and 5.13%, respectively. While it is still the case that HP pricing yields the smallest %

discount and JB pricing yields the largest % discount, the range of the % discounts are much

smaller. Furthermore, the pairwise difference in the % discount between HP and JB is no longer

significant. The data suggest that as subjects garner more experience with this negotiation

process, the effect of list price strategy on the % discount is attenuated.

Taken together, the main results of the effect of the list price strategy on the negotiation outcome

are summarized as follows:

12

Result 2a: Lower list prices lead to lower final prices.

Result 2b: In the first round of negotiations, HP pricing leads to the smallest negotiated

discount, while JB pricing leads to the largest negotiated discount.

Result 2c: Experience seems to attenuate the impact of the list price strategy on the negotiated

discount.

5.2.2 The Effect of List Price Strategy on Buyer Behavior

Because we observe all the intermediate steps within the negotiation process, we are able to also

analyze how the list price strategy separately impacts the negotiation behavior of both the buyer

and seller. We first examine the negotiation behavior of the buyer. Table 4 displays the average

buyer’s initial offer made to the seller (as a percentage of the list price) for both the first round

and in the aggregate data.

[Insert Table 4 – Effect of List Price Strategy on Buyer’s Initial Offer]

In Table 4, the average initial offer from the buyer in the first round when list prices are HP, R,

JB, and LP are 90.95%, 80.64%, 88.14%, and 87.97%, respectively. A Kruskal-Wallis strongly

rejects the null of equality of these four samples (p = .018). In addition, Mann-Whitney pairwise

comparisons reveal that the buyer’s initial offer is significantly lower when the list price is R,

compared to when it is HP, JB, and LP (p = .007, p = .072, p = .022, respectively). Furthermore,

the buyer’s initial offer is marginally higher with a HP list price compared to a JB list price (p =

.096). Overall, the data from the first round suggests buyers make the least aggressive initial

offer with a HP list price, and the most aggressive offer with a R list price.

Looking next at the aggregate sample from Table 4, the corresponding initial offers from the

buyer are 88.29%, 86.54%, 87.06%, and 88.02%, for HP, R, JB, and LP, respectively. In the

aggregate data, the same general pattern emerges but its magnitude is somewhat muted. Namely,

the buyer’s initial offer is significantly lower when the list price is R compared to HP (p = .049)

and LP (p = .089). Thus, even after some experience with the negotiation process, buyers

continue to make the most aggressive offers with a R list price.

In addition to looking at the buyer’s initial offer, we are able to analyze if, and to what extent, the

list price affects the buyer’s counter-offer behavior throughout the entire negotiation process.

Specifically, Figure 2 displays the sequence of the average buyer’s counter-offer (as a % of the

list price) for the first four offer/counter-offers.13

As expected, buyer counter-offers increase

throughout the negotiation process in all of the list price strategies. When comparing across the

different list price strategy, we observe from Figure 2 that the effect from the list price strategy

on the buyer’s initial offer does not persist throughout subsequent buyer counter-offers. That is,

for the 1st, 2

nd, and 3

rd buyer counter-offers, there are no significant pairwise differences among

any of the four different list price strategies. The data appear to suggest that while the initial list

13

We analyze only the first four offer/counter-offers because there are too few observations for any meaningful analysis after the 3

rd counter-offer. This is a result of many of the negotiations having ended in an

agreement prior to the buyer making a 4th

counter-offer.

13

price strategy does significantly affect the buyer’s initial offer, the list price strategy has very

little effect on the buyer’s subsequent counter-offers throughout the negotiation process.

The main results regarding how the list price strategy affects the negotiation behavior of the

buyer are summarized as follows:

Result 3a: In the first round of negotiations, R pricing leads to the most aggressive initial offers

by buyers, while HP pricing leads to the least aggressive initial offers.

Result 3b: Even as subjects gain experience, R pricing continues to lead to the most aggressive

initial offers by buyers, although the magnitude is smaller.

Result 3c: After the initial offer, the list price strategy has very little effect on the subsequent

counter-offers made by the buyer. That is, the list price strategy does not persistently impact

buyer behavior throughout the negotiation process.

5.2.3 The Effect of List Price Strategy on Seller Behavior

We next examine how the list price strategy affects the negotiation behavior of the seller. Table 5

displays the average seller’s 1st counter-offer made back to the buyer (as a percentage of the list

price) for both the first round and in the aggregate data. Initially considering only the first round

of negotiations, Table 5 reveals that the average 1st counter-offer made by the seller when the list

prices are HP, R, JB, and LP are 97.38%, 96.39%, 95.71%, and 94.93%, respectively. A

Kruskal-Wallis test rejects the null hypothesis of equality across these four samples (p = .026).

Furthermore, a Jonckheere-Terpstra test strongly rejects the null of equality of these percentages

in favor of the descending ordered alternative (p = .002). Thus, the lower the list price, the lower

the seller’s 1st counter-offer.

[Insert Table 5 – Effect of List Price Strategy on Seller’s 1st Counter-Offer]

Just as with buyers, when looking at the seller’s 1st counter-offer in the aggregate data, a similar

pattern emerges as in the first round, although it is less pronounced. Namely, when comparing

across the four list price strategies, an ANOVA test strongly rejects the null of equality of the

sample (p = .007). In addition, a Jonckheere-Terpstra test strongly rejects the null of equality of

these percentages in favor of the descending ordered alternative (p = .001). Thus, consistent with

the initial round results in the aggregate data, the lower the list price, the lower the seller’s 1st

counter-offer.

We next consider if, and to what extent, the list price strategy has a persistent effect on the

negotiation behavior of sellers throughout the negotiation process. To do so, we examine the

sequence of the first four average seller counter-offers across the four list price strategies. Figure

3 displays the relevant data. Unlike with buyers, the list price strategy does have a persistent

effect throughout the negotiation process. Namely, HP pricing continues to result in the highest

counter-offers, while JB pricing leads to the lowest counter-offers. The pairwise difference

between HP and JB is strongly significant for the 2nd

(p = .002), 3rd

(p = .003), and 4th

(p = .037)

counter-offers made by the seller.

[Insert Figure 3 – Sequence of Seller Counter-Offers by List Price Strategy]

14

Unlike with buyers, when sellers make their 1st counter-offer it is in response to the initial offer

made by the buyer, and the seller’s 1st counter-offer may be impacted by the buyer’s initial offer.

Hence, variation in the seller’s 1st counter-offer may be a result of variation in the buyer’s initial

counter-offer and not a response to different list price strategies. Said differently, it may be the

case that across treatments, sellers are responding to the buyers’ initial offer in a consistent way.

If this was the case, then we would expect to see sellers’ 1st counter-offers being a consistent

amount relative to buyers’ initial counter-offer across all treatments. To investigate this

possibility, we look at the effect of each list price strategy on the seller’s 1st counter-offer while

controlling for the buyer’s initial offer. Table 6 displays the results from a regression of sellers’

1st counter-offer on dummy variables representing each of the list price strategies (with R being

the excluded variable) and the buyer’s initial offer.

[Insert Table 6 – Regression Analysis of Seller’s 1st Counter-Offer]

The first column of Table 6 displays the output from a regression of the seller’s 1st counter-offer

on dummies for each of the list price strategies without controlling for the buyer’s initial offer.

Consistent with the analysis above, we see a strong monotonic relation emerge; namely, HP

pricing results in significantly higher 1st counter-offers compared to R (p = .022), JB (p < .001),

and LP (p < .001), while R pricing yields significantly higher 1st counter-offers compared to JB

(p = .004) and LP (p = .002). The second column of Table 6 displays the regression output

including the buyer’s initial offer. We can see that the buyer’s initial offer has a small but

significant positive effect on the seller’s 1st counter-offer; for every $1,000 increase in the

buyer’s initial offer, the seller’s 1st counter-offer increases by an average of $62. However, even

after controlling for the buyer’s initial offer, list price strategy still significantly impacts the

seller’s 1st counter-offer; namely, all the coefficient estimates on the dummy variable have the

same sign, similar magnitude, and remain significant.

Taken together, the main results concerning the impact of list price strategies on the negotiation

behavior of the seller are as follows:

Result 4a: In the first negotiation, there is a strong correlation between the seller’s 1st counter-

offer and the list price; the lower the list price, the lower the seller’s counter-offer.

Result 4b: As sellers gain experience with this negotiation process, lower list prices continue to

result in lower relative 1st counter-offers, although the effect is less pronounced.

Result 4c: The effect of list price strategy continues to impact the sequence of seller counter-

offers throughout the negotiation process; HP pricing always results in the highest counter-

offers, while JB pricing results in the lowest counter-offers.

Result 4d: The buyer’s initial offer has a small impact on the seller’s 1st counter-offer. Yet even

after controlling for the buyer’s initial offer, the list price strategy continues to impact the seller’s

1st counter-offer.

5.2.4 The Effect of List Price Strategy on Length of Negotiation

While the final price is an important dimension along which to evaluate the outcome of a

negotiation, another important dimension of the negotiation process is its length. That is, how

15

long does it take for the negotiating parties to reach an agreement? The length of the negotiation

process can be especially important because the resulting delays can be costly to both parties

involved.14

Because we observe the intermediate steps within the negotiation process, we are

able to explore such relations. As a proxy for the length of the negotiation process, we use the

total number of offers made by both the buyer and seller prior to an agreement being reached,

which we denote as # of total offers.15

Figure 4 compares the average # of total offers across the four list price strategies, for both the

first negotiation and when aggregated. From Figure 4, we see that in the first negotiation, there is

a strong distinction between the list price and the # of total offers. When the list price is HP, an

average of 7.06 offers are needed to reach an agreement, while this drops to 6.58 for R, 5.41 for

JB, and 4.50 when the price is LP. Statistically, a Kruskal-Wallis strongly rejects the null of

equality of these four samples (p = .008). Furthermore, a Jonckheere-Terpstra test strongly

rejects the null of equality of these percentages in favor of the descending ordered alternative (p

= .001).

[Insert Figure 4 – Comparison of Total Number of Offers by List Price Strategy]

As has been the case with much of the other analyses, the effect of the list price strategy on the

length of negotiation is attenuated when looking at the aggregated data when subjects gain

negotiation experience. Specifically, in the aggregate data, R pricing leads to the highest average

number of total offers at 8.10, while LP pricing continues to result in the least number of offers

at 7.22 (this difference is marginally significant p = .086). Although the spread within the

number of offers is much narrower across the treatments, a Jonckheere-Terpstra test still rejects

the null of equality of these percentages in favor of the descending ordered alternative (p = .045),

which provides some evidence that lower list prices do lead to shorter negotiations in the

aggregate data.

The main findings on how the list price strategy impacts the length of negotiations can be

summarized as follows:

Result 5a: In the first negotiation, the lower the initial list price, the shorter the negotiation

process (significantly fewer counter-offers are needed to reach an agreement).

Result 5b: In the aggregate data, as subjects gain negotiating experience, the effect of list price

strategy on negotiation length is diminished.

14

From the seller’s perspective, an extensive literature exists that documents the importance of mitigating “time on the market” to spare sellers carrying costs (mortgage payments, hazard insurance, and so forth). A similar argument can be made that a longer negotiation process can negatively impact buyers who expend search costs, extended stays in temporary housing, and so forth.

15 Another measure of the length of the negotiation process would be the actual time required to reach an

agreement. However, as a by-product of the experimental software, we are unable to measure time to an agreement for each negotiation pair. However, it is likely that the number of total offers needed to reach an agreement is strongly positively correlated with time, since it takes time to construct counter-offers. Thus, we argue that our results regarding how the list price affects the number of total offers would likely generalize to how the list price affects the actual time required to reach an agreement.

16

5.3 The Effect of Gender

We now explore if, and to what extent, gender of the negotiating parties acts to mediate the main

treatment effects of list price strategy on negotiation behavior.16

Specifically, we look at whether

the gender of each of negotiating party leads to a differential treatment effect. Figure 5 segments

initial offers by gender of the buyer. The result is a similar “U-shaped” pattern for both male and

female buyers. Namely, initial offers are relatively high when the price is HP, drop to their

lowest level when the price is R, and then increase again when the price is JB and LP. An

ANOVA test does not reveal a significant effect of gender (p = .144). However, this effect is

more pronounced in males, and appears to be driven by the fact that males react more than

females when the list price is LP, in the form of significantly higher average initial offers (p =

.064).

[Insert Figure 5 – Buyer’s Initial Offer by Gender and List Price]

With regard to the gender effect of the seller, Figure 6 similarly compares the seller’s 1st counter-

offer, segmented by gender. For sellers, an ANOVA test reveals a significant effect of gender on

the 1st counter-offer (p = .001). Similar to buyer behavior, this seems to be primarily driven by

the differential responses by male and female sellers to a LP list price. Compared to JB pricing,

male sellers increase their 1st counter-offer when the list price is LP, while female sellers

decrease their 1st counter-offer. Furthermore, male sellers’ 1

st counter-offer with a LP list price is

2% higher than that of female sellers, and this difference is significantly higher (p = .004).

[Insert Figure 6 – Sellers’ 1st Counter-Offer by Gender and List Price]

Taken together, the effect of gender can be summarized as follows:

Result 6a: Male and female buyers exhibit a similar pattern of initial offers across the different

list price strategies, with males making marginally higher initial offers when the list price is LP.

Result 6b: Male and female sellers exhibit a partially differential response to the list price

strategy, with males making significantly higher counter-offers when the list price is LP.

5.4 The Effect of Prior Negotiating Experience

It is likely that subjects come into the experiment with differing levels of prior negotiating

experience. This level of experience may impact how subjects react to the various list price

strategies. To explore this possibility, we collect information on each subject’s prior negotiating

experience via a post-experiment questionnaire. The question on which we construct our

measure of negotiating experience asks subjects to report how many times in the last year they

16

While there are many subject level factors that could possibly act to mediate the main treatment effects of list price strategy on negotiation behavior, we focus on gender for two main reasons. Firstly, it is easily observable in the lab. Secondly, and more importantly, there is a mature body of literature that has documented gender differences in negotiations. In lieu of citing individual studies, we refer interested readers to the following studies that provide comprehensive, although not exhaustive, views of the literature: Seagraves & Gallimore (2013) who discuss the role of gender specifically in real estate negotiations, Walters et al. (1998), Kray & Thompson (2004), and Bowles et al. (2007) who discuss gender differences in negotiations more generally, and Croson & Gneezy (2009), and Niederle & Vesterlund (2011) who review the role of gender in competition.

17

purchased an item where they negotiated the final price. We then stratify the sample into two

groups based on their answer: (i) low experience – those subjects who reported having purchased

zero goods where they negotiated the price, and (ii) high experience – those subjects who

reported having purchased five or more items where they negotiated the price.17

This

stratification roughly results in the low experience subjects representing the lowest tercile and

the high experience subjects representing the highest tercile.

Looking first at buyer behavior, Figure 7 compares the buyer’s initial offer across the list price

strategies, parsed by negotiating experience of the buyer. As one might expect, Figure 7 shows

that prior negotiating experience attenuates the effect of the list price strategy. In particular,

buyers with more experience show much less variation in their initial offers to the different list

prices. This effect is the most pronounced when comparing the initial offer associated with the

list prices HP and R. Low experience subjects offer 88.75% when the list price is HP and 85.12%

when the list price is R, while the corresponding values for the high experience buyers are

86.88% and 86.12%. Thus, the data suggest that buyers with less experience are likely to react

more, in the sense of making a lower initial offer when facing an R list price.

[Insert Figure 7 – Buyers’ Initial Offer by Negotiating Experience and List Price]

Looking next at seller behavior, Figure 8 compares the seller’s 1st counter-offer across the list

price strategies, segmented by negotiating experience of the seller. Consistent with expectations,

sellers with more negotiating experience display much less reaction in their 1st counter-offers.

Furthermore, sellers with higher levels of experience make higher or more aggressive 1st counter-

offers, regardless of the list price strategy. However, for both low and high experience sellers,

there is a consistent pattern in the effect of the list price. Namely, lower list prices tend to result

in lower 1st counter-offers by the seller, a response that is stronger for the lesser experienced

sellers.

[Insert Figure 8 – Sellers’ 1st Counter-Offer by Negotiating Experience and List Price]

The general findings regarding the effect of prior negotiation experience by the negotiating

parties on the impact of list prices and negotiation behavior is summarized as follows:

Result 7: For both buyers and sellers, higher levels of prior negotiating experience attenuates,

although not eliminates, the effect list price strategies have on initial offers and counter-offers.

6 Discussion

When examining the effect of list price strategy on final sale price, our results indicate that HP

pricing leads to the smallest negotiated discount, while JB pricing leads to the smallest

17

If we instead stratify the sample based on the median, the results are qualitatively similar, although not quantitatively as robust. In addition, we also asked subjects whether or not they had purchased an item on Craigslist (a secondary market website where transactions are most often negotiated at the time of sale), and a question of whether they had ever purchased a car (another asset where negotiation is common). If we stratify the sample into low and high experience based on whether they answered affirmatively to either of these questions, the results are also qualitatively very similar.

18

negotiated discount. In the first negotiating round, the average discount is 3.94% for HP pricing

and 5.72% for JB pricing, while in the aggregate data, the corresponding average discounts are

4.83% and 5.13%. Our results regarding the effect of list price strategies on negotiated discounts

are consistent, both in terms of direction and order of magnitude of the difference, with those

found in Beracha & Seiler (2014) who use transactions data. In addition, because of our ability,

via the controlled experiment, to exogenously vary list prices, we are able to draw causal

inference regarding the effect of list price strategy on the size of the negotiated discount. We

contend that our experimental results, combined with the transactions-based findings

documented in Beracha and Seiler, provide strong evidence that JB pricing results in the largest

discounts from the list price, while precise pricing leads to the smallest discounts.

From the seller’s perspective, what is important is maximizing the final sale price (i.e., selling

the house for the highest price, conditional on the value of the home). In our data, we find that

HP pricing leads to the highest final sale price. This finding differs from Beracha & Seiler (2014)

who conclude that JB pricing leads to the highest final prices (despite the fact that JB also leads

to the largest negotiated discounts).18

While there appears to be an inconsistency, the differing

results can be reconciled when taking into account the role of housing values. Specifically, in our

design there is no variation in home values, thus higher final sale prices associated with HP

pricing follows from the fact that the HP list price was the highest and HP pricing leads to the

smallest negotiated discounts. However, in the transactions data used by Beracha and Seiler,

there is naturally occurring variation in home values, which the authors must control for in their

study. The authors find that JB pricing is associated with the greatest degree of over-pricing.

Hence, this over-pricing outweighs the larger negotiated discounts associated with JB pricing,

which nets greater proceeds for the seller. When viewed in light of the results from Beracha and

Seiler, our findings suggest the list price strategy that maximizes the seller’s net proceeds will

depend on the extent of over-pricing associated with JB pricing. JB pricing will be preferred

when the property is over-priced, whereas HP pricing will be optimal when the list price is set

closer to the true value.

Because of our ability to observe all the intermediate steps within the negotiation process (which

is not possible with transactions data), we are able to identify how the list price strategy

separately impacts the negotiation behavior of buyers and sellers throughout the negotiation

process. Consequently, this allows us to shed light on why different list price strategies may

impact the outcome of negotiations. What we find is that buyers and seller appear to be

differentially impacted by the list price strategy. Specifically, buyers are impacted the most by R

pricing, in the form of making significantly lower initial offers. However, as the negotiation

process continues, list price has very little effect on subsequent buyer counter-offers. Conversely,

sellers are most impacted by HP and JB pricing, in the form of making higher initial counter-

offers with HP pricing and lower initial counter-offers with JB pricing. Furthermore, this effect

persists in subsequent counter-offers by the seller throughout the negotiation process. Taken

together, the results indicate that observed differences in the outcome of the negotiation based on

18

Allen & Dare (2004) also find empirical evidence just below pricing leads to higher final sales prices compared to rounded pricing, although in their paper they refer to just below pricing as “charm” pricing.

19

the list price strategy (i.e., higher final sale prices with HP pricing and larger negotiated

discounts with JB pricing) is likely attributed, in large part, to changes in the negotiation

behavior of sellers.

Another pattern that emerges from our data is the persistent difference in the negotiation process

between the HP and LP list pricing. In particular, we find that HP pricing leads to a higher final

sale price, a lower negotiated discount, and higher seller counter-offers compared to LP pricing.

Our results suggest that the position of the precise price may matter. Conditional on the seller

choosing a precise price, setting a precise price that falls just above the corresponding rounded

price may yield a higher price than choosing a precise price just below the rounded price.

However, our results also indicate that a LP list price results in a significantly shorter

negotiation, compared to a HP list price. Thus, sellers who are concerned about prolonging time-

on-the-market may opt for a LP list price because the benefit of a quick negotiation/sale with the

buyer may outweigh the disadvantage of a lower final sale price.

Throughout the analysis, a consistent result is that any observed effects of the list price strategy

in the first negotiation round data are attenuated in the aggregate data. Alternatively stated, as

buyers and sellers gain negotiating experience, the impact of the list price strategy is weakened.

This muting effect is further confirmed when we incorporate self-reported measures of

negotiating experience. Specifically, when we stratify buyers and sellers based on their

negotiating experience in other markets and analyzed the effect of list price strategy, we observe

that buyers and sellers who report being more experienced negotiators show less variation in

their counter-offers. That said, while negotiating experience did mitigate the impact of list price

strategy, it did not eliminate it. An important implication of this result is that we may expect the

effect of list price strategy to be stronger in real estate transactions by individuals, where the

negotiating parties are likely to have less experience. Yet, even when the transaction is being

brokered by experienced real estate agents, list price strategy may still impact the negotiation

process, although to a lesser extent.

Lastly, our design also enables us to explore how variation in the original purchase paid by the

seller affects negotiations, and how this may act to mediate the effect of list price strategy. Our

results indicated that, conditional on the list price, there was very little difference in seller

counter-offers when the initial purchase price was high compared to when it was low. Hence,

variation in the original purchase price likely does not lead to differential effects of the list price

strategy on negotiation. This suggests sellers do not “irrationally” respond to the manipulation of

“perceived” gains and losses based on the false reference point of the initial purchase price.

Given that the purchase price did not impact the financial payoff of the seller, we would expect

such information would not impact the seller’s behavior or the negotiation outcome, which is

what we observe in the data. This is not to say, however, that actual gains and losses would not

impact negotiations and, possibly, the effect of list price strategy. Unfortunately, because of the

challenges associated with inducing actual financial losses in an experimental setting, we were

unable to explore this possibility. Still, we view the question of how actual gains and losses

impact real estate negotiations as both interesting and important, warranting future research

efforts.

20

7 Conclusion

Buying a home is one of the most financially significant transactions a person will ever make.

One of the key decisions involved in selling a house is choosing the initial list price, from which

negotiations will typically commence. Given the order of magnitude of house prices, even small

percentage changes in the final sale price can translate into sizable absolute dollar amounts.

Therefore, it is important for the seller to choose a list price that will maximize the final sale

price and/or minimize time-on-the-market. To this end, it is imperative to first understand how

list prices impact the negotiation process and, ultimately, the outcome of the negotiation.

The main motivation of this study is to use an innovative experimental approach to investigate

the effect of initial list prices on housing negotiations. Specifically, we focus on three categories

of list price strategy – Rounded, Just Below, and Precise. We develop an experimental design

that allows us to explore if, and to what extent, these different list price strategies impact

negotiations.

Within the experiment, subjects engage in a stylized, bi-lateral housing negotiation between a

buyer and seller. We then systematically vary the initial list price strategy across treatments,

while holding the other features of the negotiating process constant. This enables us to identify

the effect of these list price strategies on the negotiation behavior of buyers and sellers, as well as

the outcome of the negotiation.

Our results indicate that the list price strategy plays an important role in the negotiations. In

terms of the negotiation outcome, we find that a high precise list price (a precise price that is just

above the rounded price) results in the highest final sale price. In terms of the negotiated

discount (percentage difference between list price and final sale price), we find that high precise

pricing leads to the smallest discount, while just below pricing leads to the largest discount. With

regards to the negotiation behavior of buyers and sellers, our results indicate that the list price

strategy impacts buyers and seller differently. Buyers make the lowest initial offer when facing a

rounded list price, but this effect dissipates in the buyer’s subsequent counter-offers throughout

the negotiation. For sellers, just below pricing results in the lowest 1st counter-offers (and is

robust to controlling for the buyer’s initial offer) and persists throughout the negotiation process.

Lastly, we find that negotiating experience tends to attenuate the above mentioned effects of the

list price strategy, although not always eliminating the effect entirely.

Our paper is not the first to examine the question of how list prices impact real estate

negotiations (see Allen & Dare, 2004; Palmon et al., 2004; and Beracha & Seiler, 2014).

However, to the best of our knowledge, we are the first to explore the question via a controlled

experiment. As such, our approach benefits from the ability to draw causal inference regarding

the effect of list price strategies on real estate negotiations, while circumventing several plausible

sources of bias associated with inferences draw from transactions-based data (e.g., systematic

selection of the list price strategy based on seller’s negotiating style/ability and/or unobservable

housing quality). Therefore, we view this study as an important and necessary complement to the

extant empirical literature aimed at better understanding how list prices impact real estate

21

negotiations. As such, the current study provides an important conceptual and methodological

contribution to the limited extant literature.

We acknowledge that there are drawbacks associated with any controlled lab experiment. Most

notably, the stylized nature of the experimental settings may limit the extent to which the results

can provide reliable inference more generally.19

That said, the controlled environment can be a

useful tool for better understanding behavior in naturally occurring environments, thus

complementing empirical and theoretical approaches. We contend that the results from

experiments are especially useful when viewed in combination with findings from related

empirical studies. In this paper, our new experimental results, along with the prior empirical

findings documented in Beracha & Seiler (2014), combine to provide a clearer and more robust

understanding of how list price strategies impact the real estate negotiation process.

When selecting a list price strategy, sellers (or their agents) may be basing their decision on

additional indirect factors other than how the list price will affect the final price directly. For

example, the choice of list price may impact the demand for the property, via search patterns of

buyers (Seiler et al. 2012), which could then indirectly impact the final sale price. Relatedly, the

choice of list price may contain signaling value (e.g., about the value of property, the seller’s

negotiating tendencies, and/or the seller’s motivation to sell), which could also indirectly impact

the final sale price. While issues of how list price strategies indirectly affect housing transactions

are beyond the scope of this study, we focus very specifically on how various list price strategies

directly impact negotiation behavior and the resulting final sale price.

We conclude by noting that real estate is by no means the only type of good where list (or

posted) price is flexible and generally viewed as negotiable. Other examples include new and

used automobiles, goods sold in secondary markets, consumer durable goods, and sub-contracted

services. We conjecture that our main results on how list price strategies impact real estate

negotiations are applicable, more generally, in these other markets where list prices are also

negotiable.

19

For a thorough discussion of topics relating to the generalizability of experiments, we refer interested readers to Levitt & List (2007). Wilde (1981) and Smith (1982) also address the topic of experimental results applying outside the lab using the term parallelism.

22

Appendix: Copy of Experimental Instructions

EXPERIMENTAL INSTRUCTIONS

PLAYER ROLE: HOME BUYER

NEGOTIATION TASK:

You have been randomly assigned to assume the role of: HOME BUYER. As the home buyer, you are

interested in purchasing a fictitious house that is for sale. Once the experiment begins, you will be

randomly paired with another participant in the lab, designated to the role of SELLER, who is interested

in selling a house to you. Your payment in the negotiation task will be based on whether you are able to

buy the house from the seller, and at what price you are able to buy the house.

Once you have been paired with the seller, you will receive the following information about this house

you are trying to buy: (i) the list price at which the house is for sale, and (ii) your reservation value for

the house, which is equivalent to the maximum price you would be willing to pay for the house.

After you receive the information about the house, you will begin the negotiation process by deciding

whether to purchase the home at the list price, or negotiate with the seller by making a counter-offer. If

you purchase at the list price, then you buy the house at the list price and the negotiation ends.

If you decide to make a counter-offer, then the seller will have two options: (i) accept your counter-

offer, or (ii) continue negotiating with you. If the seller accepts your counter-offer, then you will buy the

house at this accepted price and the negotiation ends. If the seller continues negotiating, then the seller

will make a counter-offer price back to you.

This alternating process of making price offers back and forth between you and the seller will continue

until one of the following two things happens:

i) Either you or the seller accepts the price offered by the other party.

ii) You and the seller have negotiated for the maximum allowable time of 3 minutes without either you

or the seller accepting the price offer of the other party. If you and the seller have not reached an

agreement on the price within 3 minutes, the negotiation will end, you will not buy the house, and

you will not earn any money.

EARNINGS FROM THE NEGOTIATION:

In the negotiation task, we will be using experimental currency dollars (ECDs), and all prices, offers,

counter-offers, and earnings throughout the experiment will be displayed in terms of ECDs. At the end

of the experiment, your earnings in ECDs will be converted into dollars at a rate of: $1000 ECD = $1.

Your earnings from the negotiation with the seller will depend on the following two possible outcomes

of the negotiation:

i) You buy the house: Your earnings (in ECDs) will be: $(your reservation value minus accepted price)

ii) You do not buy the house: Your earnings (in ECDs) will be: $0

23

MULTIPLE ROUNDS AND RANDOM RE-MATCHING:

You will be participating in this negotiation task a total of 4 times. In each of the 4 rounds of

negotiations, you will be randomly re-matched with a different seller. That is, the seller you will be

paired with will be different in each round. Additionally, in each round, the house’s list price, and your

reservation value may be different. However, your earnings in each of the rounds will be determined in

the same way as described above. An Experimenter will announce the conclusion of each negotiation

round. At the start of each round, the information about the new house you are potentially buying will

be displayed, so look at this information carefully before you begin.

PROCEDURE AND COMPUTER INTERFACE:

All of your decisions in each of the 4 rounds of the negotiation task will be made through the computer.

At the beginning of each round, you will see a screen that displays your reservation value for the house,

and the house’s list price. Below is an example of what you will see:

Each time the seller continues negotiating the price, you will see a screen that displays the seller’s

counter-offer. On this same screen, you will choose whether to accept the seller’s counter-offer or

continuing negotiating. To accept the seller’s counter-offer, you will click the ACCEPT button on the

lower left. To continue negotiating, you will enter your counter-offer into the box on the lower right and

click the SUBMIT button. When you are making your counter-offer, you only need to enter the numeric

value of the offer (do not enter a dollar sign or any commas). For example if you wanted to make a

counter-offer of $180,000, you would enter: 180000. Below is an example of what you will see if the

seller continues negotiating with you:

24

You and the seller will each have a maximum of 30 seconds to decide to accept the current counter-

offer or continue negotiating. A time counter is provided in the upper right corner of the screen for your

reference. As the buyer, accepting a price that is higher than your reservation value would result in you

making negative earnings; while not buying the house would result in zero earnings. Therefore, you will

not be allowed to accept any counter-offers from the seller that are more than your reservation value.

Similarly, your counter-offers to the seller must be less than or equal to your reservation value. After the

negotiation round ends, you will see a summary screen that displays the outcome of the negotiation and

the amount of earnings you generated from that round of negotiations.

EXAMPLE NEGOTIATIONS:

Here are some example negotiations to help show how your earnings are determined:

Example 1 – You receive the following information about the house:

List Price: $250,000

Your reservation value: $255,000

Suppose you do not pay the list price and you make a counter-offer of $242,500. If the seller accepts

your counter-offer:

You would receive earnings (in ECDs) of: $255,000 - $242,500 = $12,500

25

Example 2 – You receive the following information about the house:

List Price: $195,000

Your reservation value: $200,000

Suppose you did not pay the list price and you made a counter-offer of $170,500. The seller did not

accept this counter-offer and made a counter-offer back to you of $185,000. If you then accepted the

seller’s counter-offer:

You would receive earnings (in ECDs) of: $200,000 - $185,000 = $15,000

QUESTIONNAIRE:

After completing the 4 rounds of negotiations, you will be asked to complete a short questionnaire,

which will take about 5 minutes to complete. Your answers to the questions are confidential and will not

be shared with any other participants. Furthermore, your answers to the questions will not impact your

experimental earnings.

FINAL PAYMENT:

Your total experimental earnings will be determined as follows: You will receive $5 for showing up and

participating in the experiment. In addition, you will be paid your negotiation earnings from 1 of the 4

randomly selected negotiation rounds, which will be converted into dollars at a rate of $1,000 ECDs =

$1. As an example, if you made $12,000 ECDs in round 2 and that round was selected for payment, then

your total earnings would be: $5 + $12 = $17.

After you have finished the decision task and the questionnaire, you will see a final summary screen that

lists the outcome of each of the negotiation rounds, your corresponding earnings (in ECDs) for that

round, and what your final payment will be if that round is selected for payment. An Experimenter will

come to your carrel where you will role a 4-sided die to determine which negotiation round you will be

paid for. You will then be paid your total earnings that correspond to the round that was selected from

the outcome of your die role. Please remain quietly seated until the Experimenter reaches your carrel.

Once you have received your payment, please quietly exit the lab.

Thank you for your cooperation and participation!

26

EXPERIMENTAL INSTRUCTIONS

PLAYER ROLE: HOME OWNER

NEGOTIATION TASK:

You have been randomly assigned to assume the role of: HOME OWNER. As the home owner, you have

been endowed with a fictitious house that is for sale. Once the experiment begins, you will be randomly

paired with another participant in the lab, designated to the role of BUYER, who is interested in

purchasing the house from you. Your payment in the negotiation task will be based on whether you are

able to sell the house to the buyer, and at what price you are able to sell the house.

Once you have been paired with the potential buyer, you will receive the following information about

the house you are selling: (i) the list price at which the house is for sale, (ii) the original purchase price

that you paid for the home, and (iii) your reservation value for the house, which is equivalent to the

minimum price you would be willing to sell the house for.

After you receive the information about the house, the buyer will begin the negotiation process by

deciding whether to purchase the home at the list price, or make you a counter-offer. If the buyer

chooses to purchase at the list price, then you sell the house at the list price and the negotiation ends.

If the buyer decides to make you a counter-offer, then you will have two options: (i) you may accept the

buyer’s counter-offer, or (ii) you may continue negotiating with the buyer. If you accept the counter-

offer, then you will sell the house at this accepted price and the negotiation ends. If you continue

negotiating, then you will make a counter-offer back to the buyer.

This alternating process of making offers back and forth between you and the buyer will continue until

one of the following two things happens:

i) Either you or the buyer accepts the price offered by the other party.

ii) You and the buyer have negotiated for the maximum allowable time of 3 minutes without either you

or the buyer accepting the offer of the other party. If you and the buyer have not reached an agreement

on the price within 3 minutes, the negotiation will end, you will not sell the house, and you will not earn

any money.

EARNINGS FROM THE NEGOTIATION:

In the negotiation task, we will be using experimental currency dollars (ECDs), and all prices, counter-

offers, and earnings throughout the experiment will be displayed in terms of ECDs. At the end of the

experiment, your earnings in ECDs will be converted into dollars at a rate of: $1000 ECD = $1.

Your earnings from the negotiation with the buyer will depend on the following two possible outcomes

of the negotiation:

i) You sell the house: Your earnings (in ECDs) will be: $(accepted price minus your reservation value)

ii) You do not sell the house: Your earnings (in ECDs) will be: $0

27

MULTIPLE ROUNDS AND RANDOM RE-MATCHING:

You will be participating in this negotiation task a total of 4 times. In each of the 4 rounds of

negotiations, you will be randomly re-matched with a different buyer. That is, the potential buyer you

will be paired with will be different in each round. Additionally, in each round, the house’s list price, the

original purchase price, and your reservation value may be different. However, your earnings in each of

the rounds will be determined in the same way as described above. An Experimenter will announce the

conclusion of each negotiation round. At the start of each round, the information about the new house

you are selling will be displayed, so look at this information carefully before you begin.

PROCEDURE AND COMPUTER INTERFACE:

All of your decisions in each of the 4 rounds of the negotiation task will be made through the computer.

At the beginning of each round, you will see a screen that displays the original price you paid for the

house, your reservation value, and the house’s list price. Below is an example of what you will see:

Each time the buyer continues negotiating the price, you will see a screen that displays the buyer’s

counter-offer. On this same screen, you will choose whether to accept the buyer’s counter-offer or

continuing negotiating. To accept the buyer’s counter-offer, you will click the ACCEPT button on the

lower left. To continue negotiating, you will enter your counter-offer into the box on the lower right and

click the SUBMIT button. When you are making your counter-offer, you only need to enter the numeric

28

value of the offer (do not enter a dollar sign or any commas). For example if you wanted to make a

counter-offer of $180,000, you would enter: 180000. Below is an example of what you will see if the

buyer continues negotiating with you:

You and the buyer will each have a maximum of 30 seconds to decide to accept the current counter-

offer or continue negotiating. A time counter is provided in the upper right corner of the screen for your

reference. As the home owner, accepting a price that is lower than your reservation value would result

in you making negative earnings; while not selling the house would result in zero earnings. Therefore,

you will not be allowed to accept any counter-offers from the buyer that are less than your reservation

value. Similarly, your counter-offers to the buyer must be greater than or equal to your reservation

value. After the negotiation round ends, you will see a summary screen that displays the outcome of the

negotiation and the amount of earnings you generated from that round of negotiations.

EXAMPLE NEGOTIATIONS:

Here are some example negotiations to help show how your earnings are determined:

Example 1 – You receive the following information about the house:

Original Purchase Price: $240,000

List Price: $250,000

Your reservation value: $230,000

29

Suppose the buyer did not pay the list price and made you a counter-offer of $242,500. If you were to

accept the buyer’s counter-offer:

You would receive earnings (in ECDs) of: $242,500 - $230,000 = $12,500

Example 2 – You receive the following information about the house:

Original Purchase Price: $190,000

List Price: $195,000

Your reservation value: $175,000

Suppose the buyer did not pay the list price and made you a counter-offer of $170,500. You did not

accept this counter-offer (because it is less than your reservation value) and made a counter-offer of

$185,000 back to the buyer. If the buyer accepted your counter-offer:

You would receive earnings (in ECDs) of: $185,000 - $175,000 = $10,000

QUESTIONNAIRE:

After completing the 4 rounds of negotiations, you will be asked to complete a short questionnaire,

which will take about 5 minutes to complete. Your answers to the questions are confidential and will not

be shared with any other participants. Furthermore, your answers to the questions will not impact your

experimental earnings.

FINAL PAYMENT:

Your total experimental earnings will be determined as follows: You will receive $5 for showing up and

participating in the experiment. In addition, you will be paid your negotiation earnings from 1 of the 4

randomly selected negotiation rounds, which will be converted into dollars at a rate of $1,000 ECDs =

$1. As an example, if you made $12,000 ECDs in round 2 and that round was selected for payment, then

your total earnings would be: $5 + $12 = $17.

After you have finished the decision task and the questionnaire, you will see a final summary screen that

lists the outcome of each of the negotiation rounds, your corresponding earnings (in ECDs) for that

round, and what your final payment will be if that round is selected for payment. An Experimenter will

come to your carrel where you will role a 4-sided die to determine which negotiation round you will be

paid for. You will then be paid your total earnings that correspond to the round that was selected from

the outcome of your die role. Please remain quietly seated until the Experimenter reaches your carrel.

Once you have received your payment, please quietly exit the lab.

Thank you for your cooperation and participation!

30

References

Allen, M., & Dare, W. (2004). The effects of charm listing prices on house transaction prices.

Real Estate Economics 32, 695-713.

Anderson, E., & Simester, D. (2003). Effects of $9 price endings on retail sales: Evidence from

field experiments. Quantitative Marketing and Economics 1, 93-110.

Anenberg, E., & King, E. (2014). Estimates of the Size and Source of Price Declines Due to

Nearby Foreclosures. working paper, Board of Governors of the Federal Reserve System,

Washington DC.

Anenberg, E., & Laufer, S. (2014). Using Data on Seller Behavior to Forecast Short-run House

Price Changes. working paper, Board of Governors of the Federal Reserve System, Washington

DC.

Anglin, P., Rutherford, R., & Springer, T. (2003). The Trade-off Between the Selling Price of

Residential Properties and Time-on-the-Market: The Impact of Price Setting. Journal of Real

Estate Finance and Economics 26, 95-111.

Arnold, M., & Lippman, S. (1998). Posted prices versus bargaining in markets with asymmetric

information. Economic Inquiry 36, 450-457.

Basu, K. (1997). Why are so many Goods Priced to End in Nine? And Why This Practice Hurts

the Producers? Economics Letters 53, 41-44.

Basu, K. (2006). Consumer cognition and pricing in the nines in oligopolistic markets. Journal of

Economics & Management Strategy 15, 125-141.

Beracha, E., & Seiler, M. (2014). The Effect of Listing Price Strategy on Transaction Selling

Prices. Journal of Real Estate Finance and Economics, forthcoming.

Bester, H. (1993). Bargaining versus price competition in markets with quality uncertainty.

American Economic Review 83, 278-288.

Bester, H. (1994). Price commitment in search markets. Journal of Economic Behavior &

Organization 25, 109-120.

Bizer, G., & Schindler, R. (2005). Direct evidence of ending‐digit drop‐off in price information

processing. Psychology & Marketing 22, 771-783.

Bowles, H., Babcock, L., & Lai, L. (2007). Social incentives for gender differences in the

propensity to initiate negotiations: Sometimes it does hurt to ask. Organizational Behavior and

Human Decision Processes 103, 84-103.

31

Cardella, E. (2014). Negotiating Flexible Prices. Mimeo.

Chetty, R., Looney, A., & Kroft, K. (2009). Salience and Taxation: Theory and Evidence.

American Economic Review 99, 1145-1177.

Croson, R., & Gneezy, U. (2009). Gender differences in preferences. Journal of Economic

Literature 47, 448-474.

DellaVigna, S. (2009). Psychology and Economics: Evidence from the Field. Journal of

Economic Literature 47, 315-372.

Desai, P., & Purohit, D. (2004). “Let me talk to my manager”: Haggling in a competitive

environment. Marketing Science 23, 219-233.

Evans, J. & Berman, B. (1997). Marketing. NJ: Prentice Hall.

Finkelstein, A. (2009). E-ztax: Tax salience and tax rates. The Quarterly Journal of Economics

124, 969-1010.

Fischbacher, U. (2007). z-Tree: Zurich toolbox for ready-made economic experiments.

Experimental Economics 10, 171-178.

Friedman, L. (1967). Psychological pricing in the food industry. In A. Phillips & O. E.

Williamson (Eds.), Prices: Issues in theory, practice, and public policy (pp. 187–201).

Philadelphia: University of Pennsylvania Press.

Gedenk, K., & Sattler, H. (1999). The impact of price thresholds on profit contribution—should

retailers set 9-ending prices? Journal of Retailing 75, 33-57.

Gendall, P., Holdershaw, J., & Garland, R. (1997). The effect of odd pricing on demand.

European Journal of Marketing 31, 799-813.

Gill, D., & Thanassoulis, J. (2009). The impact of bargaining on markets with price takers: Too

many bargainers spoil the broth. European Economic Review 53, 658-674.

Gill, D., & Thanassoulis, J. (2013). Competition in Posted Prices with Bargaining. University of

Oxford, Department of Economics Discussion Paper Series (Ref: 639).

Han, L., & Strange, W. (2014) What is the Role of the Asking Price for a House? University of

Toronto working paper.

Haurin, D., Haurin, J., Nadauld, T., & Sanders, A. (2010). List Prices, Sale Prices, and

Marketing Time: An Application to U.S. Housing Markets. Real Estate Economics 38, 659-685.

32

Haurin, D., McGreal, S., Adair, A., Brown, L., & Webb, J. (2013). List price and sales prices of

residential properties during booms and busts, Journal of Housing Economics 22, 1-10.

Holdershaw, J., Gendall, P., & Garland, R. (1997). The widespread use of odd pricing in the

retail sector. Marketing Bulletin 8, 53-58.

Kahneman, D. (1992). Reference points, anchors, norms, and mixed feelings. Organizational

Behavior and Human Decision Processes 51, 296-312.

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk.

Econometrica 47, 263-291.

Kalyanam, K., & Shively, T. (1998). Estimating irregular pricing effects: A stochastic spline

regression approach. Journal of Marketing Research 35, 16-29.

Knight, J. (2002). Listing Price, Time on Market, and Ultimate Selling Price: Causes and Effects

of Listing Price Changes. Real Estate Economics 30, 213-237.

Korvorst, M., & Damian, M. (2008). The differential influence of decades and units on multidigit

number comparison. The Quarterly Journal of Experimental Psychology 61, 1250-1264.

Kray, L., & Thompson, L. (2004). Gender stereotypes and negotiation performance: An

examination of theory and research. In B. Staw & R. Kramer (Eds.), Research in Organizational

Behavior Vol. 26 (pp. 103-183). Oxford: Elsevier Ltd.

Kreul, L. (1982). Magic numbers: Psychological aspects of menu pricing. Cornell Hotel and

Restaurant Administration Quarterly 23, 70-75.

Kristensen, H., & Gärling, T. (2000). Anchor points, reference points, and counteroffers in

negotiations. Group Decision and Negotiation 9, 493-505.

Lacetera, N., Pope, D., & Sydnor, J. (2012). Heuristic Thinking and Limited Attention in the Car

Market. American Economic Review 102, 2206-2236.

Levitt, S., & List, J. (2007). What do Laboratory Experiments Measuring Social Preferences

Reveal About the Real World. Journal of Economic Perspectives 21, 153-174.

Miller, N., & Sklarz, M. (1987). Pricing Strategies and Residential Property Selling Prices.

Journal of Real Estate Research 2:1, 31-40.

Monroe, K. (1979). Pricing: Making Profitable Decisions. New York: McGraw-Hill.

Neale, M., & Bazerman, M. (1991). Cognition and Rationality in Negotiation. New York: Free

Press.

33

Niederle, M., & Vesterlund, L. (2011). Gender and competition. Annual Review of Economics 3,

601-630.

Palmon, O., Smith, B., & Sopranzetti, B. (2004). Clustering in real estate prices: determinants

and consequences. Journal of Real Estate Research 26, 115-136.

Poltrock, S., & Schwartz, D. (1984). Comparative judgments of multidigit numbers. Journal of

Experimental Psychology: Learning, Memory, and Cognition 10, 32-45.

Raskovich, A. (2007). Competition or collusion? Negotiating discounts off posted prices.

International Journal of Industrial Organization 25, 341-354.

Schindler, R., & Kibarian, T. (1996). Increased consumer sales response through use of 99-

ending prices. Journal of Retailing 72, 187-199.

Schindler, R., & Kirby, P. (1997). Patterns of rightmost digits used in advertised prices:

implications for nine-ending effects. Journal of Consumer Research 24, 192-201.

Schindler, R., & Wiman, A. (1989). Effects of odd pricing on price recall. Journal of Business

Research 19, 165-177.

Seagraves, P., & Gallimore, P. (2013). The Gender Gap in Real Estate Sales: Negotiation Skill of

Agent Selection? Real Estate Economics 41, 600-631.

Seiler, M., Madhavan, P., & Liechty, M. (2012). Toward an understanding of real estate

homebuyer internet search behavior: an application of ocular tracking technology. Journal of

Real Estate Research 34, 211-241.

Smith, V. (1982). Microeconomic Systems as an Experimental Science. The American Economic

Review 72, 923-955.

Stiving, M., & Winer, R. (1997). An empirical analysis of price endings using scanner data.

Journal of Consumer Research 24, 57-67.

Twedt, D. (1965). Does the ‘9 fixation’ in retail pricing really promote sales? Journal of

Marketing 29, 54-55.

Walters, A., Stuhlmacher, A., & Meyer, L. (1998). Gender and negotiator competitiveness: A

meta-analysis. Organizational Behavior and Human Decision Processes 76, 1-29.

Wang, R. (1995). Bargaining versus posted-price selling. European Economic Review 39, 1747-

1764.

Wilde, L. (1981). “On the use of Laboratory Experiments in Economics.” In P. Joseph (Ed.), The

Philosophy of Economics, pp. 137-143. Dordrecht: Reidel.

34

Table 1 – Parameter Value Combinations by Treatment

Panel A: High Purchase Price (domain of potential perceived losses for seller)

List Price Strategy

Rounded (R) $200,000 $205,000 $180,000 $195,000

Just Below (JB) $199,000 $205,000 $179,000 $195,000

Low Precise (LP) $198,674 $205,000 $178,674 $195,000

High Precise (HP) $201,326 $205,000 $181,326 $195,000

Panel B: Low Purchase Price (domain of potential perceived gains for seller)

List Price Strategy

Rounded (R) $200,000 $205,000 $180,000 $185,000

Just Below (JB) $199,000 $205,000 $179,000 $185,000

Low Precise (LP) $198,674 $205,000 $178,674 $185,000

High Precise (HP) $201,326 $205,000 $181,326 $185,000

Notes: This table summarizes the parameter values for each of the 8 treatments that comprise our experimental

design. Panel A corresponds to the high purchase price condition where the initial purchase price, denoted as

p0h, is $195,000, while Panel B corresponds to the low purchase price condition where the initial purchase price,

denoted as p0l, is set to $185,000. pl denotes the list price and its four different values correspond to the four list

price strategies we consider in the design. v and r denote the buyer’s and seller’s reservations values,

respectively in each of the treatments.

35

Table 2 – The Effect of the Purchase Price Paid by Seller

Panel A – Final Sale Price List Price Strategy

High Precise Rounded Just Below Low Precise

Initial Purchase Price ($201,326) ($200,000) ($199,000) ($198,674)

High Purchase Price 191,384 190,307 188,312 189,754

($195,000)

Low Purchase Price 191,766 190,125 189,081 187,561

($185,000)

p -value (0.613) (0.443) (0.714) (0.045)**

Panel B – Seller’s 1st Counter-Offer

High Precise Rounded Just Below Low Precise

Initial Purchase Price ($201, 326) ($200,000) ($199,000) ($198,674)

High Purchase Price 197,574 196,307 193,873 194,554

($195,000)

Low Purchase Price 197,248 195,176 194,069 192,691

($185,000)

p -value (0.374) (0.113) (0.563) (0.075)*

Notes: Panel A compares the final sale price of the house across the two different initial purchase price conditions, and Panel B compares the seller’s 1

st counter-offer across the two different initial price conditions.

Reported p-values are from a t-test of the pairwise comparison across initial purchase price conditions for a given list price strategy. All the results are robust if a Mann-Whitney U-test is used.

36

Table 3 – Effect of List Price Strategy on Negotiated Discount

% Discount

List Price Strategy

Aggregate First Round

High Precise (HP)

4.83% 3.94%

Rounded (R)

4.90% 5.12%

Just Below (JB)

5.15% 5.72%

Low Precise (LP)

5.13% 5.39%

Treatment Effects (p-values in parentheses)

Comparison of Means

(0.863) (0.309)

Trend Test (0.208) (0.076)*

HP vs R

(0.882) (0.261)

HP vs JB

(0.494) (0.056)*

HP vs LP

(0.521) (0.152)

R vs JB

(0.579) (0.579)

R vs LP

(0.611) (0.895)

JB vs LP

(0.953) (0.772)

Notes: This table compares the average negotiated discount (as a % of the initial list price) across the four list price strategies, aggregated over both initial purchase price conditions. The aggregate data comprises the data from all subjects over all four negotiating rounds, while the first round data comprises only the subset of subjects who faced the corresponding list price strategy in the first round of negotiations. Reported p-values for the comparison of means are from an ANOVA test for the aggregate data and a Kruskal-Wallis test for the first round data. Reported p-values for the trend test are from a Jonckheere-Terpstra test. Reported p-values for the pairwise comparisons are from a t-test for the aggregate data and a Mann-Whitney U-test for the first round data.

37

Table 4 – Effect of List Price Strategy on Buyers’ Initial Offer

Buyers’ Avg. Initial Offer (% of List Price)

List Price Strategy

Aggregate First Round

High Precise (HP)

88.29% 90.95%

Rounded (R)

86.54% 80.64%

Just Below (JB)

87.06% 88.14%

Low Precise (LP)

88.02% 87.97%

Treatment Effects (p-values in parentheses)

Comparison of Means

(0.348) (0.018)**

Trend Test (0.258) (0.897)

HP vs R

(0.049)** (0.007)***

HP vs JB

(0.195) (0.096)*

HP vs LP

(0.406) (0.501)

R vs JB

(0.221) (0.072)*

R vs LP

(0.089)* (0.022)**

JB vs LP

(0.278) (0.326)

Notes: This table compares the average initial offer for buyers (as a % of the initial list price) across the four list price strategies, aggregated over both initial purchase price conditions. The aggregate data comprises the data from all subjects over all four negotiating rounds, while the first round data comprises only the subset of subjects who faced the corresponding list price strategy in the first round of negotiations. Reported p-values for the comparison of means are from an ANOVA test for the aggregate data and a Kruskal-Wallis test for the first round data. Reported p-values for the trend test are from a Jonckheere-Terpstra test. Reported p-values for the pairwise comparisons are from a t-test for the aggregate data and a Mann-Whitney U-test for the first round data.

38

Table 5 – Effect of List Price Strategy on Sellers’ 1st Counter-Offer

Sellers’ Avg. 1st Counter-offer (% of List Price)

List Price Strategy

Aggregate First Round

High Precise (HP)

98.15% 97.38%

Rounded (R)

97.54% 96.39%

Just Below (JB)

96.94% 95.71%

Low Precise (LP)

96.90% 94.63%

Treatment Effects (p-values in parentheses)

Comparison of Means

(0.007)*** (0.026)**

Trend Test (0.001)*** (0.002)***

HP vs R

(0.041)** (0.207)

HP vs JB

(0.002)*** (0.052)*

HP vs LP

(0.002)*** (0.004)***

R vs JB

(0.074)* (0.792)

R vs LP

(0.075)* (0.063)*

JB vs LP

(0.467) (0.238)

Notes: This table compares the average 1st

counter-offer for sellers (as a % of the initial list price) across the four list price strategies, aggregated over both initial purchase price conditions. The aggregate data comprises the data from all subjects over all four negotiating rounds, while the first round data comprises only the subset of subjects who faced the corresponding list price strategy in the first round of negotiations. Reported p-values for the comparison of means are from an ANOVA test for the aggregate data and a Kruskal-Wallis test for the first round data. Reported p-values for the trend test are from a Jonckheere-Terpstra test. Reported p-values for the pairwise comparisons are from a t-test for the aggregate data and a Mann-Whitney U-test for the first round data.

39

Table 6 – Regression Analysis of Sellers’ 1st Counter-Offer

Seller’s 1st Counter-Offer

(1) (2)

HP 2046.77** 1713.85**

(886.37) (879.693)

JB -2652.91*** -2581.25***

(897.70) (882.81)

LP -2993.63*** -2998.93***

(929.39) (913.58)

Buyer’s Initial Offer .062***

(.022)

Constant 195451*** 184689.6***

(664.78) (3975.51)

Notes: Column (1) reports the results of an OLS regression of the seller’s 1st

counter-offer on dummy variables for each of the list price strategies (with R as the excluded variable). Column (2) reports the results of an OLS regression of the seller’s 1

st counter-offer on dummy variables for each of the list price strategies

(with R as the excluded variable) and the buyer’s initial offer. Standard errors are reported in parenthesis. p < .01 denoted by ***; p < .05 denoted by **

40

Figure 1 – Comparison of Final Sale Price by List Price Strategy

Figure 2 – Sequence of Buyer Offers by List Price Strategy

$180,000

$185,000

$190,000

$195,000

$200,000

$205,000

High Precise Rounded Just Below Low Precise

List Price Strategy

Final Sale Price

Initial List Price

Aggregate

First Round

85%

86%

87%

88%

89%

90%

91%

92%

93%

Initial Offer 1st CounterOffer

2nd CounterOffer

3rd CounterOffer

Buyer Offers and Counter-Offers

High Precise

Rounded

Just Below

Low Precise

41

Figure 3 – Sequence of Seller Counter-Offers by List Price Strategy

Figure 4 – Comparison of # of Total Offers by List Price Strategy

92%

93%

94%

95%

96%

97%

98%

99%

1st CounterOffer

2nd CounterOffer

3rd CounterOffer

4th CounterOffer

Seller Counter-Offers

High Precise

Rounded

Just Below

Low Precise

3.00

4.00

5.00

6.00

7.00

8.00

9.00

High Precise Rounded Just Below Low Precise

List Price Strategy

# of Total Offers

Aggregate

First Round

42

Figure 5 – Buyers’ Initial Offer by Gender and List Price

Figure 6 – Sellers’ 1st Counter-Offer by Gender and List Price

85.0%

85.5%

86.0%

86.5%

87.0%

87.5%

88.0%

88.5%

89.0%

89.5%

high precise rounded just below low precise

List Price Strategy

Buyers' Initial Offer (% of List Price)

Men

Women

Aggregate

95%

95%

96%

96%

97%

97%

98%

98%

99%

99%

high precise rounded just below low precise

List Price Strategy

Sellers' Initial Counter-Offer (% of List Price)

Men

Women

Aggregate

43

Figure 7 – Buyers’ Initial Offer by Negotiation Experience and List Price

Figure 8 – Sellers’ 1st Counter-Offer by Negotiation Experience and List Price

83.0%

84.0%

85.0%

86.0%

87.0%

88.0%

89.0%

90.0%

high precise rounded just below low precise

List Price Strategy

Buyers' Initial Offer (% of List Price)

Low Experience

High Experience

95.0%

95.5%

96.0%

96.5%

97.0%

97.5%

98.0%

98.5%

99.0%

high precise rounded just below low precise

List Price Strategy

Sellers' 1st Counter-Offer (% of List Price)

Low Experience

High Experience